外汇汇率时间序列静态属性的抽取

Study on Extracting Static Attributes in Time Series of Foreign Exchange Rate

为了能够使用现有的数据挖掘技术(例如粗糙集)对外汇汇率时间序列进行数据挖掘,必须从外汇汇率时间序列数据中抽取决定时间序列行为发展趋势的静态属性.针对外汇汇率时间序列的特殊性,给出了时间序列静态属性抽取技术的几个关键步骤,完成了从外汇汇率时间序列中抽取出静态属性,最后利用这些静态属性组成的数据库,实现了对外汇汇率时间序列比较准确的预测.

In order to perform existing data mining technology, such as Rough set, to mine the time series of foreign exchange rate, static attributes must be extracted to determine the developing tendency of time series behavior from the time series of foreign exchange rate. This paper presents several key steps of the time series static attributes extraction technology according to particularity of foreign exchange rate time series, then extracts static attributes from the foreign exchange rate time series, finally realizes accurate forecast to the foreign exchange rate time series by using the database composed of these static attributes.